In the 'max load' thread the question remains about how the ecuflash injector scaling affects the evoscan software and how it interprets the log data. My question is this: IF ecuflash THINKS it's telling 560 stock injectors, set to a scaling of 513, to inject a certain amount of fuel (say 20ms) to achieve an ultimate AFR of 11.5 in a given LOAD and RPM cell, and it does it and achieves the 11.5 AFR, then does that mean that the Evoscan software requires a different injector setting to read properly. Consider this, if ecuflash required 20 ms of injector time, and it was using larger injectors than it thought, 680's instead of 560's, then doesn't that mean that the ecu THOUGHT that it was using less fuel to get the job done, and therefore, less fuel means less horsepower and possible load. I understand that if ecuflash is using a higher injector scaling number (or for that matter a lower injector scaling number) than Evoscan is using, that the numbers could be skewed, but if both are the same, how can that skew the loads to interpret a HIGHER load? I understand the evoscan data.xml 5*513 settings and ipw etc. But in this case both remain the same. Does anyone know how this would skew the numbers. Again, as C6c6 said in the other thread, they should only be exactly correct for the given car, but I'm trying to understand the relationships between the 2 programs.

 

I personally stopped worrying about trying to make calculated load as correct as possible. If you haven't done the 2byte load mod, then do it now. That is the only way to know for sure what load cell you're in.

 

are you saying that the ecu calculates the load cell that it gives you in evoscan, but uses a completely different load cell algorithm for determining which target cell to use? I understand that the load cells may or may not be accurate, but the question is not the accuracy of the load, but the consistency of the look up in the maps. Then, how the difference in Ecuflash uses the calculated injector size vs how evoscan uses it. If ecuflash uses one scaling factor and evoscan uses another, I can definitely see the mathematical differences in calculation. However, if both use the same injector settings, how does that negatively affect the (correctly or incorrectly) calculated loads given in evoscan? A little complex, but the not that much. Simply put, 2 different programs, one gives the ecu the maps to work with, the other uses calculations based in part on what the ecu is given to use. It's not as simple as saying if evoscan's data.xml file isn't altered with correct injector scaling that it is ABSOLUTELY going to be incorrect and even if it is incorrect by HOW MUCH? I just thought that with so many saying the calc's were wrong, someone would surely know why and in what circumstances, such as higher settings in evoscan, lower settings in ecuflash, etc, or what if both are set to the same (but incorrectly for the size of the injectors, but none the less, the same) does that necessarily mean that the outputs will be in error. As the horsepower/torque of a motor goes up, doesn't that mean that correlate to higher load cells?

 

The load that is calculated in evoscan is an attempt to figure out what load the ECU is actually using. Its derived from a number of factors including IPW, coolant temp, intake temp, etc. Truth is, very few people, if anyone, know exactly what the ECU uses to calculate load.

Logging 2byte load, on the other hand, shows exactly what load the ECU is using. Trying to calculate load is now obsolete.

 

Thanks mismastermatt. So with that in mind, evoscan load cells may NOT be what the ecu uses (different calculation methods, possibly). The best I can relate to that then would be comparing known outputs such as LM-1 with the evoscan loads/rpms and insuring that they match. I actually have been tuning to the end result (AFR through LM-1, whp/wtq through the dyno with changes to Mivec maps according to output on the dyno, and evoscan logs according to my ignition timing maps). This is my reason for saying that the load cells are only REFERENCE points, not something that necessarily means anything other than to use in referencing the maps as closely as possible. Your response is very much appreciated. I still intend to make the changes to evoscan data.xml and see how that affects the 'displays' in previously logged files. Unfortunately, my laptop is at home and I have all of my programs and logs on it. I'll def repost with any results.

 

Correct me if I'm wrong, but the 2-byte load basically pulls a byte from two adjacent memory addresses for the load value. Evoscan then multiplies the result by 0.3125 to get the load value we see in the logs. (I don't know why it is multiplied by 0.3125, but the ECUload is 5x/8 and 0.3125 is half that.)

I have noticed that my 2-byte load is not identical to 1-byte but is within about 1% when in the 1-byte's range. This seems to be reasonable considering the slight difference in when the numbers are read. If the memory addresses are correct then the only other place where the calculation could be wrong is the multiplication by 0.3125. (Forgetting about latency.) So, if evoscan is reading the right numbers they should be pretty accurate although not perfect because of varying read times.

 

First test done. Didn't expect to see any difference, but just covering ALL bases. Previous logs made using Evoscan with injectors set to 513 AND ecuflash using 513, then I changed evoscan settings to 609, repulled loggraphs up and even though evoscan's CURRENT settings are 609, the logs made with 513 settings display absolutely no differences in loads. I didn't think it would show a difference, but just needed to verify before going to next step of logging both ways, 513 evoscan/609 evoscan, and verify the differences in display.